Even for patients who have maintained long-term abstinence from drugs, relapse remains a substantial risk. As we have shown using ecological momentary assessment (EMA), lapses to drug use may follow acute increases in stress. In the rat reinstatement model of relapse, stress-induced seeking of heroin, cocaine, speedball (heroin-cocaine combination), alcohol, or nicotine is blocked by alpha-2 adrenoceptor agonists such as lofexidine, guanfacine, and clonidine. Thus, alpha-2 agonists may act on a final common pathway of stress-induced relapse, relevant to multiple drugs of abuse. In a randomized, placebo-controlled laboratory study, with non-treatment-seeking cocaine users, we have shown that clonidine was effective in reducing stress-induced (and, at a higher dose, cue-induced) craving in a pattern consistent with the findings from the reinstatement model. We have now taken that finding much further having just completed a successful clinical trial and thereby identified a major new use for clonidine. We also incorporated EMA as an ourcome measure to test hypotheses about clonidines behavioral mechanism of action. In our randomized, double-blind, placebo-controlled trial, we enrolled 208 opioid-dependent outpatients at our buprenorphine clinic. Participants (118, 57%) who maintained abstinence continuously for 2 weeks were continued on buprenorphine and randomized to clonidine or placebo for 14 weeks. Urine was tested thrice weekly. Lapse was defined as any opioid-positive or missed urine; relapse, as consecutive lapses. Time to lapse and relapse were examined with Cox regressions; longest period of abstinence, with a t-test; EMA data, with generalized linear mixed models. In an intent-to-treat analysis, clonidine significantly increased the longest duration of consecutive days of abstinence for opioids during intervention. There was no group difference in time to relapse, but the clonidine group took significantly longer to lapse. EMA self-reports collected in the participants natural environments showed that daily-life stress was partly decoupled from opioid craving in the clonidine group, supporting our hypothesized mechanism for clonidines benefits. Thus, clonidine is useful in opioid dependence not just for alleviation of withdrawal signs, but also as an adjunct maintenance treatment that increases duration of abstinence: even in the absence of physical withdrawal, it decouples stress from craving in everyday life. We also conducted a trial with an almost identical rationale and design, except that the target was prevention of relapse to cocaine rather than opioids. The medication we tested was the partial D2 agonist aripiprazole, because it had been shown to block reinstatement of cocaine-seeking behavior maintained by either cocaine-paired cues or by a priming injection of cocaine. We had planned to randomize 110 participants. Unfortunately, out of the first 40 enrolled, only 45% met the abstinence criterion of two continuous weeks of abstinence for randomization. In an interim analysis, the aripiprazole group (15 mg/day) seemed to have more cocaine-negative urines than the placebo group, but the differences were not statistically significant. In both groups, all but 3 participants relapsed. There was no difference between the groups in time to relapse. Due to slow enrollment and low rates of achievement of initial abstinence, low rate of randomization, and the unpromising initial results, we ended the trial. The lack of success in this trial contrasts starkly with those obtained in the clonidine trial. We are also evaluating the role of stress in relapse in a large natural-history study in which real-time field monitoring of stressor exposure is combined with continuous location tracking via GPS. Preliminary analyses suggest some unexpected relationships between neighborhood environment and self-reported stress. As we collect more data, we should be able to determine how patterns of environmental-stressor exposure predict relapse. One of our goals is to supplement our ambulatory assessments with on-the-spot feedback, turning them into mobile interventions.